Problem: Advertising is hard, reaching the correct people with advertising is hard -- probably easier with Adwords than in the real world, but still hard --, and is getting harder as more money is spent on internet advertising. At the same time, ad blockers are quickly spreading and maybe will soon kill all the sources of advertising, bad for personal websites, bad for small companies using Adwords.
Solution: Think affiliate marketing. Now think manual advertising contracts (when a blogger or a podcaster explicitly recommends some website or local business in his writing/talking). This tool would combine these two to get some level of automation in the process. It would beat ad blockers, give a personal touch to the advertising and release companies and bloggers from the pain of manually contracting.
How would it work?
(Let's imagine this service is hosted at natural.com, which is obviously not true). Let's see an example. A blogger writes about the topic of fishing for beginner fishers. He has a lot of fisher-wannabe readers. He talks about things the wannabe fishers must buy and places they must go. Every time he talks about a product, he puts a link to natural.com/, or when he talks about a place, he can use something like natural.com/, maybe he can also create more complex link patterns, but let's stick to these for now. These links are going to redirect the clicker.
Advertisers would bid for these pre-placed links. Or, in other words, when there's a link to natural.com/fishing-rod, it can redirect to some fishing rod shop that is paying, and the blogger will get money for each click on that link. The service could use some intelligent criteria to decide which fishing rod shop would appear at each time -- for example, geolocation. When there's a link to natural.com/galapagos, any business from Galapagos could apply, but perhaps the context of the link -- a fishing blog -- would influence the decision on which Galapagos business would actually appear. Tourism companies could also appear depending on the geolocation and so on. These are all open and complex criteria that should be decided, tested and experimented.
When there's no company eligible for appearing at the link (when I say "appear" I mean getting the link visitor to be redirected to the company website), a static fallback could be specified by the blogger, or the service could just redirect the visitor to a Google search on the topic -- initially I thought about the service doing an intelligent guess and redirecting the user accordingly, but now it seems like a stupid thing to do since Google does a job so enormously better in figuring out these things at the moment.
Other markets where this could clearly work: food blogs; programming blogs; travel blogs; hack-your-life sites; personal development blogs; clothing and fashion blogs; any blog or site about hobbies or practices like yoga, painting, martial arts, hunting etc.; all podcasts and radio shows where the podcaster tells the listeners to go to an address like mypodcast.com/fish-offers; probably many others I'm forgetting at the time I write.
Progress: No progress at all. I realize the network-effect of this thing is damn hard to beat, but at the same time it is something that I see working in small scale, so it is probably doable.
Market: Bloggers; small business; local, non-tech small businesses. Later, big companies that sell a lot of different things, big content websites (newspapers, magazines etc.).
Team: One developer with many ideas, this one being the only one hard enough to be worth posting here.